作者位次 #3/8ACL 2026Poster· Natural Language Processing
E ye M ulator: Improving Code Language Models by Mimicking Human Visual Attention
Yifan Zhang · Chen Huang · Yueke Zhang · Jiahao Zhang · Toby Jia-Jun Li · Collin McMillan · Kevin Leach · Yu Huang
摘要
Code Language Models (CodeLLMs) traditionally learn attention based solely on statistical input-output token correlations (“machine attention”). In contrast, human developers rely on intuition, selectively fixating on semantically salient tokens during program comprehension. We present EyeMulator...